Accelerating multiscale finite element simulations of history-dependent materials using a recurrent neural network

Accelerating multiscale finite element simulations of history-dependent materials using a recurrent neural network

Ghavamian, F. (TU Delft Applied Mechanics)
Simone, A. (TU Delft Applied Mechanics; Università degli Studi di Padova)

2019

FE² multiscale simulations of history-dependent materials are accelerated by means of a recurrent neural network (RNN) surrogate for the history-dependent micro level response. We propose a simple strategy to efficiently collect stress–strain data from the micro model, and we modify the RNN model such that it resembles a nonlinear finite element analysis procedure during training. We then implement the trained RNN model in the FE² scheme and employ automatic differentiation to compute the consistent tangent. The exceptional performance of the proposed model is demonstrated through a number of academic examples using strain-softening Perzyna viscoplasticity as the nonlinear material model at the micro level.

Deep learning
Machine learning
Multiscale modeling
Recurrent neural network
Strain softening
Viscoplasticity

http://resolver.tudelft.nl/uuid:3832b270-d25c-4c06-8a96-000d62298d9d

https://doi.org/10.1016/j.cma.2019.112594

2021-08-23

0045-7825

Computer Methods in Applied Mechanics and Engineering, 357

Accepted author manuscript

Institutional Repository

journal article

© 2019 F. Ghavamian, A. Simone